e258533d529cec3da5e03c8c0c023d7d

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the nyu-mll/glue [stsb] dataset. It achieves the following results on the evaluation set:

• Loss: 2.5525
• Data Size: 1.0
• Epoch Runtime: 35.3504
• Mse: 0.6384
• Mae: 0.6257
• R2: 0.7144

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Mse	Mae	R2
No log	0	0	56.5548	0	3.5986	14.1395	2.9673	-5.3251
No log	1	179	98.0311	0.0078	3.9008	24.5076	4.0091	-9.9631
No log	2	358	78.0175	0.0156	4.2950	19.5045	3.9340	-7.7251
No log	3	537	6.2264	0.0312	5.4949	1.5574	1.0349	0.3033
No log	4	716	4.7232	0.0625	7.2611	1.1811	0.8730	0.4717
No log	5	895	5.2693	0.125	9.7318	1.3177	0.9495	0.4105
1.5559	6	1074	3.6177	0.25	13.7798	0.9046	0.7498	0.5953
3.1937	7	1253	3.8436	0.5	21.8007	0.9614	0.7984	0.5699
3.48	8.0	1432	2.4446	1.0	37.9666	0.6114	0.6248	0.7265
2.0803	9.0	1611	2.5939	1.0	35.3437	0.6486	0.6312	0.7099
1.4826	10.0	1790	2.5311	1.0	34.4862	0.6328	0.6188	0.7169
1.197	11.0	1969	3.0228	1.0	34.9335	0.7560	0.6939	0.6618
1.0154	12.0	2148	2.5525	1.0	35.3504	0.6384	0.6257	0.7144

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1